Accommodative Stability in the Developing Visual System

Purpose: The purpose of this work was to gain an understanding of the impact that extrinsic, intrinsic and cognitive factors have on accommodative accuracy (lag) and variability (RMS) in young children with uncorrected hyperopia. Individuals with greater amounts of hyperopia must exert a larger accommodative response to achieve the same level of accommodative accuracy as an individual with a lesser hyperopic refractive error. Given that accommodative responses are positively correlated with RMS, which in turn is positively correlated with depth-of-field (DOF), children with uncorrected hyperopia may experience greater RMS that may also impact their ability to detect blur. Additionally, the accommodative response has been shown to be influenced by the cognitive demand of the viewing task, which may impact the RMS observed in these children. Experiments investigating accommodative accuracy and variability under different viewing conditions were performed on both children with varying amounts of uncorrected hyperopia and adults for comparison with a mature visual system.

Methods: Lag and RMS were measured over 3 experiments using photorefraction in children 3 to <10 years with a range of uncorrected hyperopia (+0.06D to +4.91D spherical equivalent (SE) of the most plus eye) and visually normal uncorrected adults (+0.08D to +1.51D SE). Experiment 1 investigated the impact blur and disparity cues have on lag and RMS by systematically removing these cues from the stimulus. Relationships between total accommodative response and lag and RMS were investigated. Experiment 2 measured blur detection thresholds and DOF psychophysically. Associations between blur detection thresholds with independent variables RMS, age, SE and lag and associations between DOF and RMS were investigated. Experiment 3 investigated the impact of cognitive effort on lag and RMS as functions of SE and the total accommodative response for both passive and active viewing conditions.

Results: Experiment 1) In children, accommodation was most accurate and stable when blur and disparity cues were present in the stimulus. The removal of blur cues from the stimulus resulted in both greater lag and RMS than the removal of disparity cues. In adults, RMS was not impacted by the removal of blur or disparity while lag was significantly greater when blur-cues were removed. Experiment 2) In children, increased RMS, increased SE and younger age were significantly associated with increased blur detection thresholds. RMS and age were independently associated with blur detection thresholds when controlling for the other independent variables. Additionally, increased SE was significantly associated with increased RMS when controlling for age and lag. Experiment 3) In children, increased cognitive effort resulted in a smaller average accommodative lag and decreased RMS. Increased SE was significantly associated with increased lag and increased RMS in the passive viewing condition only. In adults, cognitive effort did not significantly affect lag, however, a small significant difference in RMS was detected with cognitive effort compared to passive viewing but the difference is unlikely to be clinically meaningful. Children vs. Adults) The children had significantly greater DOF and blur detection thresholds and RMS than the adult subjects throughout all experiments. Conclusions: Children do not have adult-like RMS, which may be secondary to children having larger blur detection thresholds and DOF than adults. Children appear to rely more heavily on both blur and disparity cues for accommodative accuracy and stability than adults. Additionally, cognitive effort appears to improve accommodative accuracy and stability in children more than adults, as measured responses in adults did not change with cognitive effort. Lastly, the impact of uncorrected hyperopia on lag and RMS was variable across conditions suggesting that while children with uncorrected hyperopia may experience greater lag and RMS at times, their accommodative behavior is not consistently poorer across near tasks.